Setup of Arduino nano sense to test out Sensor Fusion

Question/Issue: I cannot get the serial daemon to work. Also the command $ edge-impulse-daemon comes back with an error that the command cannot be recognised.

**Project ID:**This is before I get into a project - I am just setting up

Context/Use case: This is before I get into a project - I am just setting up. I would also like to explore the “real time” browser based interface before I get into as serious project.
PLEASE help. I have installed node.js tools as someone suggested - but no change.
I can see the device thought he Arduino CLI which reports it is on com port 3 and have installed the “updated firmware” via the process indicated by @janjongboom videos which report it has found the device on Comm Port 4 Any help from anyone would be greatly appreciated.

Hello @odsouza,

Can you copy the error output from the edge-impulse-daemon please?
Also can you see your device using the WebUSB (from the Data Acquisition tab in your Edge Impulse project)?

If not, maybe the default edge impulse firmware has not been uploaded properly onto the board. Which OS are you using?

Also, I think from the Arduino Nano 33 BLE Sense, you need to press the reset button on the board before using the daemon.

Let me know if any of these hints help.

Regards,

Louis

Thank you very Louis for your reply (we had our compulsory federal elections yesterday hence could not reply. I guess something is not right because the point at which I am stuck is the Edge Impulse Daemon installation - which seemed so straightforward. I followed Jan’s instructions and setup the Arduini CLI - works ok and provides me the status of the device on Com3. I then proceeded to upload the firmware from the repository suggested - and this too uploaded OK and reported the device is on Com4. I tried setting up the serial daemon - but this did not connect to my device. I am stuck at this piont.
Microsoft Windows [Version 10.0.19043.1706]
(c) Microsoft Corporation. All rights reserved.

  • I checked if Python was to be running in the background (this did not help so I shut it down)
  • I checked if NodeJS was a requirement - by installing it. (this did not help)
    See transcript of my interaction with the CLI environment - none of the commands suggested in the tutorial worked after I installed the firmware…
    See transcript below:
    C:\Users\Ollencio D’Souza>node.js
    ‘node.js’ is not recognized as an internal or external command,
    operable program or batch file.

C:\Users\Ollencio D’Souza>edge-impulse-daemon
‘edge-impulse-daemon’ is not recognized as an internal or external command,
operable program or batch file.

C:\Users\Ollencio D’Souza> edge-impulse-daemon
‘edge-impulse-daemon’ is not recognized as an internal or external command,
operable program or batch file.

C:\Users\Ollencio D’Souza>$ edge-impulse-daemon
‘$’ is not recognized as an internal or external command,
operable program or batch file.

C:\Users\Ollencio D’Souza>edge-impulse-

I would appreciate some help to sort these basic issues out (sorry - for my ignorance - but I am catching up all on my own at 1am in the morning!)

One more thing Louis - I tried to install the firmware again, twice - and this worked both times - no errors.

Hi Louis - it all worked well over the web interface. Thanks for your help.
One question I need to ask is if there is a multiple fusion requirement- has anyone done a combined setup of the NN so it can take in all these inputs and look for anomalies in the combined data reaching the sensors. I have been through Jan’s tutorials - just superb - where he discusses key word and other aspects. Has he done on a “fusion” test case?
Thanks and
Regards

Can you type the following command which node to see if it is located somewhere? Did you make sure to install the Additional tools when installing NodeJS: Installation - Edge Impulse Documentation

has anyone done a combined setup of the NN so it can take in all these inputs and look for anomalies in the combined data reaching the sensors

Sure, you can add a second learning block set as anomaly detection and you will be able to see the anomaly score and you classifier score during the inference: Continuous motion recognition - Edge Impulse Documentation

Regards,

Louis